bert_12_layer_model_v1 / train_results.json
gokuls's picture
End of training
758391f
raw
history blame contribute delete
196 Bytes
{
"epoch": 5.0,
"train_loss": 3.446043330256489,
"train_runtime": 56381.9546,
"train_samples": 2929379,
"train_samples_per_second": 259.78,
"train_steps_per_second": 2.03
}